How to Connect Google Search Console to Gemini
Quick answer: Gemini reads Google Search Console through an MCP server in the Gemini CLI. Add the server URL under mcpServers in settings.json, restart the CLI, and complete the sign-in to pick which properties it may read. The consumer Gemini app does not support custom MCP connectors as of publication, so use the CLI or a CSV upload.
It seems like the one assistant that should read Search Console natively is Google's own. Yet you can ask the Gemini app about your property all day and get generic SEO advice, because the app cannot see your data any more than a stranger can.
The working path runs through the Gemini CLI, Google's open-source terminal agent, which supports MCP servers the same way Claude, Cursor, and Codex do. This guide covers that setup, what each Gemini surface supports as of publication, and the honest fallback when a terminal is out of the question.
SEARCH CONSOLE
How to Connect Google Search Console to Gemini
PRIMARY KEYWORD
connect Google Search Console to Gemini
READ TIME
8 minutes
Which Gemini surfaces support MCP as of publication?
The Gemini CLI supports MCP servers through its settings file, and that support is documented in the project itself. The consumer Gemini app and website connect to Google's own services, such as Workspace apps, and do not offer custom MCP connectors as of publication. Developers can also wire MCP servers into the Gemini API in their own applications, which is a separate build-it-yourself path.
Google ships AI surfaces quickly and renames them often, so treat this map as dated the day it was written. The Gemini CLI documentation linked below is the source to trust over this paragraph.
How do you add a Search Console MCP server to the Gemini CLI?
The CLI reads MCP configuration from a settings.json file, globally in your home directory's .gemini folder or per project. Add the Search Console server, such as the free read-only SEO Receipts endpoint, under the mcpServers key with its URL.
On the next start, the CLI discovers the server's tools. Complete the sign-in it presents, choose which Search Console properties to share, and the tools become available in your sessions. The /mcp command lists connected servers and their tools, which is the quickest way to confirm the connection took.
- Open or create settings.json in the .gemini folder
- Add an mcpServers entry whose URL points at the MCP server
- Restart the CLI and run /mcp to confirm the server is listed
- Complete the OAuth sign-in and select the properties to share
- Ask a test question whose answer you already know
What can Gemini answer once connected?
The same bounded analysis any MCP-connected assistant handles: top queries and pages for a finalized window, comparisons between two date ranges, the queries behind one URL, striking-distance candidates, and sitemap status. Scope every question with a property, a date range, and a threshold.
The CLI's natural home is development work, so the highest-value questions are the ones that precede a change: what a page earns before you delete it, which queries a URL holds before you retitle it, whether the sitemap read cleanly after a deploy.
What if you use the Gemini app instead of the CLI?
Export and upload. Filter the Performance report to the rows that matter, export the CSV, and give it to the app with a scoped prompt. The analysis will be frozen at export time and capped by the interface's 1,000-row export limit, which is fine for a single page review and cramped for a sitewide one.
Disclose those limits in anything you write up afterward. An export-based analysis covers the exported rows only, excludes the queries Google withholds for privacy, and cannot answer follow-ups outside its columns.
How do you keep the connection under control?
The rules are identical to every other assistant. Use a server whose tools are read-only, share only the properties the work needs, and prefer project-level configuration when client data is involved so one repo cannot see another client's property.
Revocation should be a two-layer affair: remove the server from settings.json to cut the CLI's access, and revoke the server's Google grant from your Google account's third-party access page when you are done with the service entirely.
TAKEAWAY CHECK
What belongs on the connect gemini to search console cleanly?
- 01Choose a read-only Search Console MCP server.
- 02Add it under mcpServers in the Gemini CLI settings.json.
- 03Restart, run /mcp, and complete the property-scoped sign-in.
- 04Verify one known number against the Performance report.
- 05Keep client properties in per-project configuration.
- 06Remove the settings entry and revoke the Google grant when finished.
What else do people ask about connect Google Search Console to Gemini?
Why can't the Gemini app just read my Search Console account?
Google runs them as separate products with separate permissions. The app's connections cover services like Workspace, and Search Console is not among them as of publication. Nothing about owning both accounts grants the assistant access, which is arguably the correct default for data this sensitive.
Is the Gemini CLI free to use?
Google offers free usage tiers with sign-in, plus paid options through API keys, and the limits change often enough that the project's documentation is the only current source. The MCP server side can be free independently, as the SEO Receipts server is in public beta.
Does this give Gemini access to my whole Google account?
No. The MCP server's OAuth grant covers the scopes it requests, which for a read-only Search Console server means viewing search data for the properties you select. The grant is separate from your Gemini sign-in and can be revoked on its own.
Can I use the same MCP server with Gemini and Claude at once?
Yes. An MCP server is a service, and each client connects to it independently with its own authorization. Analysts commonly point Claude, Cursor, and the Gemini CLI at the same read-only endpoint, and each connection is granted and revoked separately.
The CLI lists the server but shows no tools. What now?
Usually the sign-in never completed. Re-run the authentication for that server, finishing the browser flow in the same session, and confirm the authorized Google account actually has access to the property in Search Console. Then restart the CLI and check /mcp again.
Which primary sources support this guide?
Product behavior and metric definitions change. These are the official Google references used for this article and checked on July 28, 2026.
- 01Google: Gemini CLI
Google's open-source Gemini CLI, including MCP server configuration in settings.json.
- 02Model Context Protocol documentation
The open specification for connecting AI assistants to external tools and data sources.
- 03About Search Console data
Google's documentation on freshness, privacy omissions, row limits, time zones, and discrepancies.
- 04Google Account Help: third-party apps and services with account access
Google's instructions for reviewing and removing third-party access to a Google account.
FREE GOOGLE SEARCH CONSOLE MCP SERVER
Ask your AI assistant about your own Search Console data.
Connect the free SEO Receipts MCP server once. Claude, ChatGPT, Cursor, or Codex can then read clicks, impressions, queries, pages, and sitemap status from the properties you select.